1 Introduction

The XBRL Query and Rules Language is a specialized query language designed for querying XBRL (eXtensible Business Reporting Language) data and uses a data model based on the XBRL Open Information Model (OIM). Its primary purpose is to effectively query XBRL data by extracting data points, performing calculations and applying rules or validations. The notable feature of XBRL Query and Rules Language is that it is syntax-independent and can support processing queries of XBRL data in any supported format (e.g., xBRL-XML, xBRL-CSV, xBRL-JSON).

1.1 Relationship to other work

XBRL Query and Rules Language is a successor to XBRL Formula 1.0.

XBRL Query and Rules Language builds upon the capabilities of XBRL Formula 1.0 and expands its capabilities to also provide broader querying capabilities. This enhancement supports the growing complexity and diversity of use cases for XBRL data. XBRL Query and Rules Language provides more flexibility and functionality beyond the scope of providing calculations and validation.

1.2 Terminology

The key words MUST, MUST NOT, REQUIRED, SHALL, SHALL NOT, SHOULD, SHOULD NOT, RECOMMENDED, MAY, and OPTIONAL, in this specification, are to be interpreted as described in IETF RFC 2119.

The keywords expanded name, NCName and QName are to be interpreted as described in the XML Names specification.

1.3 Documentation conventions

QNames in parenthetical red text after a "MUST" or "MUST NOT" statement prescribe standardized error codes to be used if the preceding condition is violated. "MUST" or "MUST NOT" statements that do not have a prescribed error code are not automatically enforceable, and processors are not required to detect violations.

1.4 Namespaces and prefixes

This specification makes use of QName notation to refer to XML expanded names (the combination of a namespace URI and a local name, and the xBRL-JSON format make use of abbreviated forms for these identifiers. The prefixes used by this specification are shown in the table below. These prefixes are reserved prefixes.

The prefix dtr-type denotes any namespace that is the namespace for a type defined in the Data Types Registry.

2 Querying XBRL Data (Informative)

The XBRL Query and Rules Language is a domain specific language used to define queries and assertions over an XBRL instance or taxonomy. The language is based on the XBRL data model. Queries are written in text files. A query is basically a formula that can access XBRL data and calculate a new value. The calculated value can either be outputted or used to generate an assertion message.

The result of the query is a message or set of messages as output. The message can be either the query result or can be an assertion. Assertions are used to test XBRL data and produce messages when the assertion is not met. For example, Equities is not equal to Assets + Liabilities is an assertion that when the sum of the values for Assets and Liabilities does not equal Equities a message is produced to indicate the data is out of balance.

output EquitiesValue
    @Assets + @Liabilities

assert OutOfBalance
    @Equities#e != @Assets#a + @Liabilities#l
message
    'The value of Equities {$e} does not equal the sum {$a + $l} of Assets {$a} and Liabilities {$l}'

The language has two distinct components: fact filtering and taxonomy navigation.

2.1 Querying for XBRL Facts

Every fact defined in an XBRL instance document has dimensions associated with it, that define the fact. These dimensions include the built in dimensions:

• The reporting concept of the fact

• The period the fact of time at which the fact is measured

• The entity for which the fact is reported

• The unit of measurement (for numeric facts)

In addition to the built in dimensions, facts can have additional taxonomy dimension, such as "Business Segment Dimension", "Geography Dimension", and "Energy Source Dimension". These dimensions allow disaggregated facts to be reported.

Instance document filtering allows queries to use these dimensions to select facts. XBRL Query and Rules Language provides 2 special features when filtering instance documents.

• Creating iterations of the query for each occurrence of a fact

• Aligning facts with in an iteration of the query

2.1.1 Creating Query Iterations

A unique feature of XBRL Query and Rules Language is that when multiple facts are selected to use in an expression, such as to list values for Revenues, XBRL Query and Rules Language will automatically repeat the execution of the query for each occurrence of the fact in the XBRL data.

output EquitiesValue
    @Revenue
message "Revenues: {$rule-value} for period: {$rule-value.period}"

Revenues: 10,000 for period: 2021-01-01 to 2021-12-31
Revenues: 20,000 for period: 2022-01-01 to 2022-12-31
Revenues: 30,000 for period: 2023-01-01 to 2023-12-13

2.1.2 Alignment of facts

When combining multiple facts in a query the XBRL Query and Rules Language will align the facts based on the dimensions of facts. For example, if calculating the value of Equities by adding Assets and Liabilities, only the Assets and liabilities for the same period would be calculated.

output EquitiesValue
    @Assets#a + @liabilities#l
message "{$rule-value} (Equities) = {$a} (Assets) + {$l} (Liabilities) for period: {alignment()['period']}"

210,000 (Equities) = 100,000 (Assets) + 110,000 (Liabilities) for period: 2020-12-31
410,000 (Equities) = 200,000 (Assets) + 210,000 (Liabilities) for period: 2021-12-31
610,000 (Equities) = 300,000 (Assets) + 310,000 (Liabilities) for period: 2023-12-31

Note that the Assets and liabilities are automatically aligned based on period. By default, the XBRL Query and Rules Language will not create an iteration where the Assets and liabilities have different periods. How the processor aligns facts in an iteration is controllable by the query.

2.2 Taxonomy navigation

In addition to query for XBRL facts, the XBRL Query and Rules Language provides extensive support for navigating taxonomies. For example, the XBRL Query and Rules Language can compare the structure of a company extension taxonomy against the underlying authoritative taxonomy (i.e. US GAAP Taxonomy or IFRS Taxonomy).The resulting taxonomy relationship sets can then be combined with a fact query to determine where values have been used. These navigation features are explained in detail in the Navigation section of this document. (see Section 7)

2.3 Data Model

The query language follows the XBRL data model, defining objects and properties that can be used to understand an XBRL filing or Taxonomy. The model operates at a semantic level - no operations are dependent on any underlying xml, json or csv syntax, although the model does allow access to some syntactic components using functions. This document defines all of the object classes and properties of those classes. An understanding of the XBRL data model is essential to use the query language.

3 Query Data Model

The Query Data Model is composed of the Taxonomy Object Model and the Instance Object Model.

An object represents the components of XBRL taxonomies and XBRL instances. An object represents a unique XBRL artifact such as a concept or a label. All objects are comprised of properties that define the attributes of the object.

An object property defines the attributes of an object.

The query language operates on objects defined in the Query Data Model.

The Query Data Model does not support tuples.

The Taxonomy Object Model defines the objects in a valid XBRL taxonomy. The Taxonomy Object model includes properties of each object.

The taxonomy object model shown in Figure 1 is comprised of the following objects:

Figure 1: Taxonomy Object Model

The Instance Object Model defines the objects in an XBRL Instance. The Instance Object model includes properties of each object.

The instance object model shown in Figure 2 is comprised of the following objects:

Figure 2: Instance Object Model

Properties of each object are accessed using a dot notation.

$cube.cube-concept.data-type

3.1 Taxonomy Objects

The taxonomy object model is comprised of the following components.

3.1.1 Taxonomy Object

The taxonomy object defines the properties of a taxonomy.

constant $US-GAAP-2020 = taxonomy('https://xbrl.fasb.org/us-gaap/2020/entire/us-gaap-entryPoint-all-2020-01-31.xsd')

3.1.1.1 Taxonomy Object Properties

The taxonomy concepts property returns a set of all concept objects comprising the taxonomy object.

$US-GAAP-2020.concepts

The taxonomy concept-names property returns a set of all concept QNames comprising the taxonomy object.

$US-GAAP-2020.concept-names

The taxonomy cubes property returns a set of all cube objects comprising the taxonomy object.

$US-GAAP-2020.cubes

The taxonomy dimensions property returns a set of all dimension objects comprising the taxonomy object.

$US-GAAP-2020.dimensions

The taxonomy dimensions-explicit property returns a set of all explicit dimension objects comprising the taxonomy object.

$US-GAAP-2020.dimensions-explicit

The taxonomy dimensions-typed property returns a set of all typed dimension objects comprising the taxonomy object.

$US-GAAP-2020.dimensions-typed

The taxonomy dts-document-locations property returns a set of url or file paths comprising all the files that make up the discoverable taxonomy set (dts).

$US-GAAP-2020.dts-document-locations

The taxonomy entry-point property returns the url of the entry point file of the taxonomy object.

$US-GAAP-2020.entry-point

The taxonomy entry-point-namespace returns the uri of the namespace for the entry point file of the taxonomy object.

$US-GAAP-2020.entry-point-namespace

The taxonomy networks property returns a set of all network objects comprising the taxonomy object.

$US-GAAP-2020.networks

The taxonomy namespaces property returns a set of all namespace uri's comprising the taxonomy object. This includes every namespace comprising the dts.

$US-GAAP-2020.namespaces

The taxonomy roles property returns a set of all role objects comprising the taxonomy object. This does not include arcroles.

$US-GAAP-2020.roles

The taxonomy arcroles property returns a set of all arcrole objects comprising the taxonomy object.

$US-GAAP-2020.arcroles

The taxonomy part-elements property returns a set of all part-element QNames comprising the taxonomy object.

$US-GAAP-2020.part-elements

The taxonomy elements property returns a set of element QNames that are not part-elements or concepts comprising the taxonomy object.

$US-GAAP-2020.elements

The taxonomy data-types property returns a set of all type objects comprising the taxonomy object.

$US-GAAP-2020.data-types

The taxonomy uri property returns the url of the entry point file of the taxonomy object.

$US-GAAP-2020.uri

3.1.1.2 Taxonomy Object Properties/Functions with Parameters

The taxonomy object has the following functions.

• concept(concept qname ) (see Section 11.2.1)

• cube(statement table concept|QName, drs-role)(see Section 11.2.2)

• effective-weight(concept1, concept2) (see Section 11.2.3)

• effective-weight-network(concept1, concept2) (see Section 11.2.4)

• element(custom name) (see Section 11.2.5)

• networks(arcrole, extended link role) (see Section 11.2.6)

3.1.2 Concept Object

The concept object defines an XBRL concept. The concept object includes the concepts properties and traits. The query specification treats labels and references as properties of the concept object.

constant $Assets = $US-GAAP-2020.concept(Assets)

3.1.2.1 Concept Object Properties

The concept balance property returns a string value of either debit, credit or empty string. This property corresponds to balance defined in the XBRL specification.

$Assets.balance

The concept base-type property returns a base-type of the data-type of the concept. The base-type is represented as a type object.

$Assets.base-type

The concept data-type property returns a data-type of the concept. The data-type is represented as a type object.

$Assets.data-type

The concept enumerations property returns a set of strings derived from the data-type object of the concept. This is a short form of $concept.data-type.enumerations

$Assets.enumerations

The concept ext-enum property returns a set of enumerated values allowed for the concept based on an ext-enum object.

$Assets.ext-enum 

The concept has-enumerations property returns a boolean derived from the data-type object of the concept. This is a short form of $concept.data-type.has-enumerations

$Assets.has-enumerations

The concept is-abstract property returns a boolean indicating the concept is an abstract, a value of true or the concept is not an abstract, a value of false.

$Assets.is-abstract 

The concept is-monetary property returns a boolean indicating if the QName of the data type object of the concept equals xbrli:monetaryItemType.

$Assets.is-monetary

The concept is-numeric property returns a boolean value of true indicating if the QName of the base-type is derived from decimal, double or float.

$Assets.is-numeric

The concept local-name property returns a string representing the suffix of the QName.

$Assets.local-name

The concept name property returns a QName of the concept.

$Assets.name

The concept namespace-uri property returns a uri representing the namespace of the QName of the concept.

$Assets.namespace-uri

The concept nillable property returns a boolean indicating the concept is nillable.

$Assets.nillable

The concept period-type property returns a string value of either instant or duration string. The period-type attribute when defined in a expression it is defined as a keyword.

$Assets.period-type

The concept all-references property returns a set of all reference objects associated with the concept.

$Assets.all-references

The concept substitution property returns a type object of the substitution group of the concept.

$Assets.substitution

The concept traits property returns a set of trait concept objects associated with the concept.

$Assets.traits

3.1.2.2 Concept Object Properties/Functions with Parameters

The concept object has the following associated properties that take parameters.

• all-labels(label role, language) (see Section 11.3.1)

• attribute(attribute name)(see Section 11.3.2)

• is-type(data type) (see Section 11.3.3)

• label(label role, language) (see Section 11.3.4)

• references(reference role) (see Section 11.3.5)

3.1.3 Cube Object

The cube object defines the properties of an XBRL hypercube based on the drs role.

constant $CUBE = $US-GAAP-2020.cube(ScheduleOfFairValueOffBalanceSheetRisksTable, FinancialInstrumentsFairValueDisclosuresScheduleOfFairValueOffBalanceSheetRisksTable)

3.1.3.1 Cube Object Properties

The cube cube-concept property returns a concept object or hypercube concept in the substitution group xbrldt:hypercubeItem

$CUBE.cube-concept

The cube closed property returns a boolean value representing if the hypercube is closed, a value of true or the hypercube is open, a value of false.

$CUBE.closed

The cube drs-role property returns a role object representing the drs role of the cube.

$CUBE.drs-role

The cube dimensions property returns a set of dimension objects of the cube. Each dimension is a dimension object.

$CUBE.dimensions

The cube dimensions-explicit property returns a set of explicit dimension objects of the cube. Each dimension is a dimension object.

$CUBE.dimensions-explicit

The cube dimensions-typed property returns a set of typed dimension objects of the cube. Each dimension is a dimension object.

$CUBE.dimensions-typed

The cube primary-concepts property returns a set of primary concepts of the cube. Each primary concept is represented as a concept object. Primary concepts do not include domain members.

$CUBE.primary-concepts

The cube facts property returns a list of fact objects that are valid in the cube. The cube includes nil and duplicate facts.

$CUBE.facts

3.1.3.2 Cube Object Properties/Functions with Parameters

The cube object has the following associated properties that take parameters.

• dimension(dimension QName) (see Section 11.6.1)

3.1.4 Dimension Object

The dimension Object defines the properties of an XBRL dimension based on the drs role. The dimension object is different from the dimension concept. The dimension object is a dimension on a cube in a given link role with members, domains and defaults. The dimension concept is a concept object representing the properties of the concept. The dimension concept is the dimension concept property of the dimension object.

constant $DIMENSION = $CUBE.dimension(srt:StatementGeographicalAxis)

3.1.4.1 Dimension Object Properties

The dimension concept property returns the concept object for the dimension.

$DIMENSION.concept

The dimension dimension-type property returns a string of either typed or explicit

$DIMENSION.dimension-type

The dimension default property returns the concept object for the default concept of the dimension.

$DIMENSION.default

The dimension members property returns a set of concept objects for the domain members of the dimension.

$DIMENSION.members

The dimension cube property returns the cube object for the dimension object.

$DIMENSION.cube

3.1.5 Relationship Object

The relationship object defines an XBRL relationship between two concepts. The relationship object does not include relationships between concepts and resources. The relationship object includes generic relationships.

constant $NETWORK_BS = networks($US-GAAP-2020,parent-child,'http://fasb.org/us-gaap/role/statement/StatementOfFinancialPositionClassified')

constant $PRES_RELATIONSHIP_OBJECT = first(filter sum($NETWORK_BS).relationships where $item.source-name == NotesAndLoansReceivableNetNoncurrentAbstract and $item.target-name == NotesAndLoansReceivableGrossNoncurrent)

3.1.5.1 Relationship Object Properties

The relationship source property returns a concept object that represents the source concept in a relationship object.

$PRES_RELATIONSHIP_OBJECT.source

The relationship source-name property returns a QName that represents the source concept in a relationship object.

$PRES_RELATIONSHIP_OBJECT.source-name

The relationship target property returns a concept object that represents the target concept in a relationship object.

$PRES_RELATIONSHIP_OBJECT.target

The relationship target-name property returns a QName that represents the target concept in a relationship object.

$PRES_RELATIONSHIP_OBJECT.target-name

The relationship order property returns a decimal that represents the order attribute of the relationship object.

$PRES_RELATIONSHIP_OBJECT.order

The relationship weight property returns an integer that represents the weight attribute of the relationship object.

$PRES_RELATIONSHIP_OBJECT.weight

The relationship preferred-label property returns a uri that represents the preferred label of a relationship object.

$PRES_RELATIONSHIP_OBJECT.preferred-label

The relationship role property returns a role object that represents the extended link role the relationship object is present in.

$PRES_RELATIONSHIP_OBJECT.role

The relationship arcrole property returns a role object that represents the arcrole of the relationship object.

$PRES_RELATIONSHIP_OBJECT.arcrole

The relationship arcrole-uri property returns a string that represents the uri of the arcrole of the relationship object.

$PRES_RELATIONSHIP_OBJECT.arcrole-uri

The relationship arcrole-description property returns a string that represents the description of the arcrole of the relationship object.

$PRES_RELATIONSHIP_OBJECT.arcrole-description

The relationship link-name property returns a QName that represents the linkbase the relationship object is defined in.

$PRES_RELATIONSHIP_OBJECT.link-name

The relationship arc-name property returns a QName that represents the arc-name the relationship object is defined in.

$PRES_RELATIONSHIP_OBJECT.arc-name

The relationship network property returns a network object that represents the network the relationship object is present in.

$PRES_RELATIONSHIP_OBJECT.network

3.1.6 Label Object

The label object defines the properties of an XBRL label. If the label object is used as a property of a concept then the label object will default to the standard label. To get all labels the all-labels function of the concept is used. (see Section 11.3.1

3.1.6.1 Label Object Properties

The label lang property returns the language of a label as a string

$US-GAAP-2020.concept(Assets).label.lang

The label role property returns the role object of the label.

$US-GAAP-2020.concept(Assets).label.role

The label text property returns the text of a label as a string

$US-GAAP-2020.concept(Assets).label.text

3.1.7 Reference Object

The reference object defines the properties of an XBRL reference.

3.1.7.1 Reference Object Properties

The reference parts property returns a set of part objects.

The reference role property returns the role object of the reference.

3.1.8 Reference-part Object

The reference-part object defines the properties of an XBRL reference part.

3.1.8.1 Reference-part Object Properties

The reference-part part-value property returns a string representing the value of the part.

The reference-part name property returns the QName of the part. The property follows the format $part.name where $part is the reference reference-part object.

The reference-part namespace-uri property returns a uri representing the namespace of the reference-part name property. The property follows the format $part.namespace-uri where $part is the reference reference-part object.

The reference-part local-name property returns a string representing the local name of the reference-part name property. The property follows the format $part.local-name where $part is the reference reference-part object.

The reference-part order property returns a decimal representing the order of the part in the reference > The property follows the format $part.order where $part is the reference reference-part object. The part order property is not defined in the XBRL specification. The part order property is calculated based on the order of the parts in the taxonomy schema document.

The reference-part element property returns the QName of the part element.

 $con = $US-GAAP.concept(Assets);

for ($ref in $con.references('http://www.xbrl.org/2003/role/presentationRef'))
    list("reference\n",
        for ($p in $ref.parts)
         $p.name.string + " - " + $p.part-value + "\n"
    ).join("")

reference
ref:Publisher - FASB
ref:Name - Accounting Standards Codification
codification-part:Topic - 942
codification-part:SubTopic - 210
ref:Section - S99
ref:Paragraph - 1
ref:Subparagraph - (SX 210.9-03(11))
codification-part:URI - http://asc.fasb.org/extlink&oid=6876686&loc=d3e534808-122878

3.1.9 Network Object

The network object represents a single network in the taxonomy, identified by its extended link role, arcrole, standard arc element, and standard extended link element .

constant $NETWORK_BS = $US-GAAP-2020.networks(parent-child,'http://fasb.org/us-gaap/role/statement/StatementOfFinancialPositionClassified')

3.1.9.1 Network Object Properties

The network arcrole property returns a role object that represents the arcrole of the network object.

$NETWORK_BS.arcrole

The network arcrole-description property returns a string that represents the description of the arcrole of the network object.

$NETWORK_BS.arcrole-description

The network arcrole-uri property returns a string that represents the uri of the arcrole of the network object.

$NETWORK_BS.arcrole-uri

The network concept-names property returns a set of concept QNames that are included as source-name and target-name of the relationships in the network object.

$NETWORK_BS.concept-names

The network concepts property returns a set of concept objects that are included as source and target concepts of the relationships in the network object.

$NETWORK_BS.concepts

The network source-concepts property returns a set of concept objects that are included as source concepts of the relationships in the network object.

$NETWORK_BS.source-concepts

The network target-concepts property returns a set of concept objects that are included as target concepts of the relationships in the network object.

$NETWORK_BS.target-concepts

The network relationships property returns a set of relationship objects in the network object.

$NETWORK_BS.relationships

The network role property returns the extended link role object of the network object.

$NETWORK_BS.role

The network roots property returns a set of concepts representing the root concepts of the network object.

$NETWORK_BS.roots

3.1.10 Role Object

The role object is used to define an XBRL role.

constant $NETWORK_BS_ROLE = $NETWORK_BS.role

3.1.10.1 Role Object Properties

The role uri property returns the role uri as a string.

$NETWORK_BS_ROLE.uri

The role description property returns a string that represents the description of the role object.

$NETWORK_BS_ROLE.description

The role used-on property returns a set of QNames of the extended link element names the role is used on.

$NETWORK_BS_ROLE.used-on 

The role all-references property returns a set of all reference objects associated with the role.

$NETWORK_BS_ROLE.all-references

3.1.10.2 Role Object Properties/Functions with Parameters

The role object has the following associated properties that take parameters.

• all-labels(label role, language) (see Section 11.3.1)

• label(label role, language) (see Section 11.3.4)

• references(reference role) (see Section 11.3.5)

3.1.11 Ext-enum Object

The ext-enum object is used to define an extensible enumeration.

3.1.11.1 Ext-enum Object Properties

The ext-num role property returns the role object of the ext-num object.

The ext-num domain property returns the concept object of the ext-num domain.

The ext-num members property returns a set of concepts linked to the domain concept by the domain-member arcrole.

The ext-num head-useable property returns a boolean that indicates if the domain member is included as an enumeration.

The ext-num enum-version property returns a string representing the version of extensible enumeration used.

3.1.12 Type Object

The taxonomy model uses the XML Schema data-types system [XML Schema Datatypes] for defining the datatype of values. It does not make use of the XML Schema structures system [XML Schema Structures], as this is not relevant outside of the context of an XML document.

A type object represents a data type.

3.1.12.1 Type Object Properties

The type name property of a type object returns the QName of the type, for simple types.

The type enumerations property of a type object returns a set of allowed values for the type.

The type has-enumerations property of a type object returns a boolean if the type is restricted to a list of enumerated values.

The type-ancestry property of a type object returns a list of the data type ancestry.

The parent-type property of a type object returns the type of the parent type.

The base-type property of a type object returns the base type of the type.

The min-inclusive property of a type object returns an integer value of the min inclusive cardinal value for a type.

The max-inclusive property of a type object returns an integer value of the max inclusive cardinal value for a type.

The min-exclusive property of a type object returns an integer value of the min exclusive cardinal value for a type.

The max-exclusive property of a type object returns an integer value of the max exclusive cardinal value for a type.

The total-digits property of a type object returns the integer value of the total digits of a value.

The fraction-digits property of a type object returns an integer value of the number of digits to the right of the decimal place.

The value-length property of a type object returns an integer value of the length of a string value.

The min-length property of a type object returns an integer value of the minimum length of a string value.

The max-length property of a type object returns an integer value of the maximum length of a string value.

The white-space property of a type object returns a string of one of the following: 'preserve', 'replace' or 'collapse'.

The pattern property of a type object returns a single regex expression or a list/set of regex expressions.

3.1.13 Name Object

The name object represents a QName.

3.1.13.1 Name Object Properties

The QName namespace-uri property of a name object returns the uri of the QName as a string.

The QName local-name property of a name object returns the local name of the QName as a string.

3.2 Instance Object Model

The instance object model is comprised of the following components.

3.2.1 Instance Object

The instance object defines the properties of a instance.

constant $WSFS = instance('https://www.sec.gov/edgar/wsfs-20211231.htm')

constant $CURRENT_INSTANCE = instance()

3.2.1.1 Instance Object Properties

The instance document-location property returns the document location of the instance object as a uri.

The instance facts property returns a list of facts in an instance. These facts do not have dimensional alignment. The facts property returns all facts including duplicates. This differs from the fact query which removes duplicates.

The instance taxonomy property returns the taxonomy object used by the instance.

3.2.2 Fact Object

The fact object defines a fact. The value of the object defaults to the fact value as a formatted string when output. There is no property that represents the fact value.

3.2.2.1 Fact Object Properties

The fact decimal property returns an integer representing the decimals associated ith a numeric fact. If the fact is a non-numeric fact a none value is returned.

The fact concept property returns the concept object for the fact.

The fact period property returns the period object for the fact. If the fact has a forever period then a none value is returned.

The fact unit property returns the unit object for the fact. If the fact has non-numeric value a none value is returned.

The fact language property returns the language string for the fact. If the fact has a numeric value or no language assigned a none value is returned.

The fact entity property returns the entity object for the fact. If the fact has no entity property a none value is returned.

The fact dimensions property returns a dimensions dictionary associated with the fact. If the fact has no dimensions a none value is returned. The dimensions dictionary is a dictionary object with key value pairs. The dictionary key must be either a dimension concept or QName. The dictionary value must be either a dimension member concept, QName or datatype if a typed dimension.

The fact dimensions-explicit property returns a dimensions dictionary associated with the fact if the dimensions are explicit. If the fact has no explicit dimensions a none value is returned.

The fact dimensions-typed property returns a dimensions dictionary associated with the fact if the dimensions are typed. If the fact has no typed dimensions a none value is returned.

The fact id property returns the fact id as a string. If the fact has no id a none value is returned.

The fact sid property returns the sid associated with the fact. A sid is a semantic identifier based on the concept, unit, period, entity, dimensional qualifications, decimals value and the value of the fact. The sid is defined using a hash. The calculation of the sid hash is processor dependent!

The fact footnotes property returns a set of footnote objects. If the fact has no footnotes an empty set is returned.

The fact instance property returns the instance object that the fact is in.

The fact cubes property returns a set of cubes that the fact is valid in. If the fact is not valid in a cube an empty set is returned.

The fact is-fact property returns a boolean result if the object is a fact object.

The fact is-monetary property returns a boolean result if the fact has a monetary value.

The fact is-nil property returns a boolean result if the fact has a nil value.

The fact namespace-map property returns a set of namespaces defined in the instance document.

3.2.2.2 Fact Object Properties/Functions with Parameters

The fact object has the following associated property that takes parameters.

• dimension(dimension qname) (see Section 11.4.1)

3.2.2.3 Inline Fact Object Properties

Inline fact properties are included as properties of the fact for facts defined in an inline document. The value of the inline fact properties will return none when an XBRL instance is not in an inline XBRL format.

The fact inline-hidden returns a boolean result if the fact is hidden in the inline document.

The fact inline-scale returns an integer result defining the scale used in the inline document.

The fact inline-format returns a QName result defining the transform that was used for the fact in the inline document. If no transform was used a none value is returned.

The fact inline-display-value returns a string result representing the string value appearing in the inline document.

The fact inline-negated returns a boolean result if the fact is negated in the inline document.

The fact inline-parents returns a list of facts that are parent html nodes of the fact in the inline document. The property navigates continuations in the inline document. If a fact is included in multiple continuations the parents of those continuations are included in the list of parent facts.

The fact inline-ancestors returns a list of facts that are ancestor html nodes of the fact in the inline document. The property navigates continuations in the inline document. If a fact is included in multiple continuations the ancestors of those continuations are included in the list of ancestor facts. When going up the tree, continuations are evaluated for each parent fact and each parent continuation. If a fact is in a continuation that was referenced by another continuation the parent is accessed by traversing back to the source continuation and evaluating parent html nodes of that continuation.

The fact inline-children returns a list of facts that are child html nodes of the fact in the inline document. The property navigates continuations in the inline document. If a fact includes continuations the children of those continuations are included in the list of child facts.

The fact inline-descendants return a list of facts that are descendant html nodes of the fact in the inline document. The property navigates continuations in the inline document. If a fact includes multiple continuations the descendants of those continuations are included in the list of descendant facts. When going down the tree continuations are evaluated for each child fact and each child continuation.

3.2.2.4 Fact through Concept Object Shortcut Properties

The fact all-references property returns a set of all reference objects associated with the concept of the fact.

The fact base-type property returns a base-type of the data-type of the concept associated with the fact. The base-type is represented as a type object.

The fact data-type property returns a data-type of the concept associated with the fact. The data-type is represented as a type object.

The fact enumerations property returns a set of strings derived from the data-type object of the concept associated with the fact. This is a short form of $concept.data-type.enumerations

The fact has-enumerations property returns a boolean derived from the data-type object of the concept associated with the fact. This is a short form of $concept.data-type.has-enumerations

The fact is-numeric property returns a boolean indicating if the QName of the base-type of the concept associated with the fact equals decimalItemType.

The fact name property returns a QName of the concept associated with the fact.

The fact local-name property returns a string representing the localName of the QName.

The fact namespace-uri property returns a string representing the prefix of the QName.

The fact substitution property returns a type object of the substitution group of the concept associated with the fact.

3.2.3 Period Object

The period object defines the period information associated with a fact value in the instance.

3.2.3.1 Period Object Properties

The period days property of a period object returns the number of days in a duration as an integer. Days are calculated where dates are the start of the day. The days property for an instant period will return 0 days.

The period end property of a period object returns the end date of a duration period or the date time of an instant as a date type

The period start property of a period object returns the start date of a duration period or the date time of an instant as a date type.

3.2.4 Duration Object

The duration object represents a start date and end date duration. A duration object can be defined using the duration function.

3.2.4.1 Duration Object Properties

The duration days property of an duration object returns the number of days in a duration as an integer.

The duration end property of an duration object returns the end date time of the duration as a date type.

The duration start property of an duration object returns the start date time of the duration as a date type.

3.2.5 Instant Object

The instant object represents a specific date time. An instance object can be defined using the date function.

3.2.5.1 Instant Object Properties

The instant days property of an instant object returns 0 days.

The instant end property of an instant object returns the date time of the instant as a date type.

The instant start property of an instant object returns the date time of the instant as a date type.

3.2.6 Unit Object

The unit object defines the unit of a fact value in the instance. Only numeric facts can have a unit.

3.2.6.1 Unit Object Properties

The unit denominator property of a unit object returns the xbrl measure of the denominator as a QName. If there is no denominator an empty value is returned.

The unit numerator property of a unit object returns the xbrl measure of the numerator as a QName. If the unit only has a measure and no division the property returns the measure.

The unit id property of a unit object returns the id of the unit used in the instance.

3.2.7 Entity Object

The entity object defines the entity of a fact value in the instance.

3.2.7.1 Entity Object Properties

The entity id property of an entity object returns the id of the entity used in the instance.

The entity scheme property of an entity object returns the scheme of the entity used in the instance.

3.2.8 Footnote Object

The footnote object defines the footnote associated with a fact in the instance.

3.2.8.1 Footnote Object Properties

The footnote arcrole property of a footnote object returns the arc-role of the footnote.

$footnote.arcrole

The footnote content property of a footnote object returns the content of the footnote. If it is a fact it returns a fact object

The footnote fact property of a footnote object returns a set of facts the footnote relates to.

The footnote lang property of a footnote object returns the language of the footnote.

The footnote role property of a footnote object returns the footnote-role of the footnote.

$footnote.role.uri

4 Query Set

A query set represents all the physical components used to query either a taxonomy or instance documents.

The query set includes the following components:

1. Namespace Declarations
2. Namespace Group Declarations
3. Output Attribute Declarations
4. Function Declarations
5. Constants
6. Message Default Language
7. Assert Declarations
8. Output Declarations
9. Query Name Prefix
10. Query Name Separator

The query set can be defined in one file or be spread across multiple files. The first five components above are global. For example a constant or custom function defined in one file is available to Query blocks defined in a different file.

The last two components defined above are file specific components that streamline the naming of queries.

• Query name prefix declaration
• Query name separator declaration

The files contained with a query set must be defined within a common directory. All the files used to define the query set must be identified using the suffix .xql.

4.1 Namespace Declarations

Namespace declarations are used in order to determine the namespace of QNames. Namespace prefixes are used when defining QNames. The prefixes and the associated namespace uri's are defined in the query set in the following format.

namespace {prefix}={uri}

Namespaces are defined once in the query set and cannot be duplicated.

namespace iso4217 = http://www.xbrl.org/2003/iso4217

The default namespace is declared by excluding the prefix.

namespace http://fasb.org/us-gaap/2024

The namespace uri is not defined as a string.

The namespace prefix must be defined as a simple name.

The namespace prefix must be unique. xbrlqe:NamespacePrefixNotUnique

4.2 Namespace Group Declarations

In some cases it is unknown what the namespace of an element will be. XBRL Query expressions expect to be executed against a given taxonomy, but when queries are run against extension taxonomies unexpected namespaces can arise when different versions of the same taxonomy are used. The namespace group averts this by using QName prefixes that can represent a group of namespaces.

The namespace group declaration maps a single namespace prefix to a list of defined namespaces.

The group of namespaces is provided as a list of string items. A string item MUST be provided in the list. xbrlqe:InvalidObjectType

The string items provided in the list can be a full uri or partial portion of the uri. The processor should match all namespaces that match the full or a partial component of the string.

Multiple namespace groups can be defined in a query set. Each namespace group prefix MUST be unique.xbrlqe:NamespacePrefixNotUnique

The namespace group declaration eliminates the risk of duplicate element name clashes across different taxonomies when local-names are used. It allows a single query set to be defined that can be used across updated releases of a taxonomy which has a time incremented namespace.

    namespace dei = http://xbrl.sec.gov/dei/2021
    namespace dei-2022 = http://xbrl.sec.gov/dei/2022

    namespace-group dei-all = list('dei')

    {@concept = dei-all:DocumentPeriodEndDate}

    {@concept.local-name = 'DocumentPeriodEndDate'}

Multiple namespaces can be included in the list of allowable namespaces by providing them as strings in the list.

A namespace group MUST not be declared in the fact query. NamespaceGroupDeclarationInFactQuery

A namespace group prefix MUST only be used in a fact query expression. NamespaceGroupPrefixNotInFactQuery

4.3 Output Attribute Definitions

The query set can contain a user defined Output Attribute that is associated with the output of a query. This is useful for classifying rule types for applications. Output attributes allow for refinement on severity levels or additional details that would otherwise have to be extracted from a message string.

Output attributes are defined using the output-attribute declaration as shown below:

output-attribute myQueryAttribute

The output attribute defines an attribute called myQueryAttribute. This can then be defined at the bottom of a query block.

The output attribute is unique and duplicate definitions must not be defined. If a duplicate is defined the error xbrlqe:DuplicateOutputAttributeDefined is returned.

Predefined output attributes are included called message, query-suffix, query-focus and severity that are not defined using an output attribute declaration.

In addition to predefined output attributes, the query language includes output attributes that supports exporting information to a file. The attribute file-content defines the content that will be output. This is defined as a string and can be derived from the to-json, to-csv, to-spreadsheet or string properties. The content MUST not contain objects.

The output attribute file-location defines the location where the file content is written to. Multiple output queries can be written to the same file. To prevent the existing content from being overwritten the output attribute file-append can be used with a boolean value to indicate if content is appended or overwritten. The value of file-append MUST be true or false.

output ppe-spreadsheet

      $spreadsheet = dict(list('PPE', list('Property Net',[@PropertyPlantAndEquipmentNet]))).to-spreadsheet

      true
else
    skip

file-location '/output/ppe.xlsx'
file-content $spreadsheet
file-append true

4.4 Custom Function Definitions

A Custom function can be used when queries repeat the same logic. Rather than duplicating the query logic, a custom function can be defined and used by multiple queries. Custom functions allow the passing of values to the function and returning the results from the function. Variables or tags defined within the function are available to use in a query output such as a message. Functions are defined with the keyword function.

Function names must be defined as a simple name.

Dimensional alignment of facts is maintained when using a function.

Recursive functions are not supported by the processor.

Functions do not have to have arguments.

function non_neg_concepts()
set(Assets, Liabilities)

The railroad diagram in Figure 3 shows the components of an Custom function declaration.

Figure 3: Function Declaration

4.5 Constants

A constant is a global variable defined as part of the query set but not included in the query block.

• A constant cannot be defined within a query block.

• A constants type is not declared.

• A constant is assigned using the = operator.

• A constant takes the type of the assigned value or object.

• A constants name is prefixed with a $ sign.

• A constants value cannot be reassigned by a constant with the same name.

• A constant name is prefixed with the keyword constant.

• A constant name must be a simple name.

• Values recorded in a constant are iterable.

• A constant can contain a query expression.

• A constants name is case sensitive.

• The keyword constant is case-insensitive.

constant $hello = 'hello'

4.6 Message Default Language

The message-language keyword is used to identify the default language used for messages. If no message language is defined it defaults to English. The message language can be specifically be defined as part of the message.

message-language "fr"

4.7 Query Blocks

Queries are defined in blocks called a query block. Query blocks have a specific structure depending on the query type. A query can either be defined as an assert query or as an output query.

Assert queries are used to define a rule and must evaluate to a boolean result. The keyword assert is used to identify a query as an assertion.

The railroad diagram in Figure 4 shows the components of an assert declaration.

Figure 4: Assert Declaration

Output queries are used to return data from an XBRL instance, XBRL taxonomy or both. The keyword output is used to identify a query as an output.

The railroad diagram in Figure 5 shows the components of an output declaration.

Figure 5: Output Declaration

4.7.1 Parts of a Query Block

Query Parts are those components that make up a query block.

Assert and output queries are composed of the following parts.

4.7.1.1 Query Declaration

The query declaration part of a query block must use either the keyword assert or output. These keywords are case insensitive. The query declaration is the first part of the query. The declaration indicates if the query is an assert query or an output query.

4.7.1.2 Query Name

The query name is the second part of the query block. The query name must be defined as a NCname. The query name must be unique across the entire query set. Query names are often called rule names.

4.7.1.3 Query Assertion Type

The assertion type is the third part of the query block. The assertion type can have a value of satisfied or unsatisfied or be blank. The assertion type must be blank for an output query. If not defined for an assert query the default is satisfied. The satisfied keyword returns a result if the expression is satisfied. The unsatisfied keyword returns a result if the expression is not satisfied

4.7.1.4 Query Expressions

Query Expressions are comprised of local variables, tagged values, block expressions, fact queries, navigate queries, properties, functions, literals and operators to produce a query result. For an assert query the query expressions must return a true or false result. An output query is not restricted to the type of the value returned. A query expression can use any of the grammar defined in the query specification.

4.7.1.5 Query Result

The results of a query is called a query result. The query result can have five possible components. The railroad diagram in Figure 6 shows the components of a query result.

QueryResult

Figure 6: Query Result

4.7.1.5.1 Query Message

The query message is defined using the keyword message. The query message can be any expression that is converted to string when output to report information from either an assert or output query. The query message cannot create iterations. All variables calculated in the query expression can be inserted into a message string using a curly bracket substitution.

A message is required for an assert query. A message is not required for an output query.

Within a query expression tagged values can be defined using the # character. The name of the tag must succeed the # character.

{@concept = assets}#a

The value and properties of # tags can be referenced in the query message. Any values in a query expression can be tagged. This includes the contents of custom functions.

Tagged values, local variable names and expressions can be inserted into message strings using the curly bracket substitution syntax. Any expression can be defined in curly brackets within a message to report local variable values or tagged values that were defined as part of the iteration.

output AssetValue
    {@concept = Assets}#a 

message
    "The value of {$a.concept.name} is {$a}"

The value of the query expression is recorded in a variable called $rule-value. If a rule evaluates to a boolean of true, then the value of $rule-value will be recorded as 'true'. The $rule-value variable can be used in a message to indicate the result of an output or the result of running the rule.

output add-two-numbers
    @assets#a + @liabilities#b

message
    "{$a} +  {$b} = {$rule-value}"

4.7.1.5.1.1 Message Language

The default language used for messages is set using the message-language keyword. To output messages in a different language a language string is defined after the message keyword using the following format:

message "BCP 47 language code"

The string MUST be a valid BCP 47 BCP47 language code oime:invalidLanguage.

    message "fr"  

    "message en français"

The language code can be evaluated as an expression that resolves to a string.

Multiple messages can be defined in different languages.

4.7.1.5.2 Query Severity

The severity part is defined using the severity keyword. The severity keyword is case insensitive Severity can be used with the assert and output query. The severity is used to indicate the severity of a rule result. Severity must have the following enumerated values: error, warning, ok or pass. The severity can also be defined as a variable or an expression that resolves to a string with a value of error, warning, ok or pass. The allowable enumerated values are case insensitive. If a severity message other than error, warning, ok or pass is defined then an error of xbrlqe:InvalidSeverityValue is returned.

assert myerror satisfied

$error = 'warning'
true

message
"This is an error."

message "fr"
 "C'est une erreur"

error $error

4.7.1.5.3 Query Focus

The query-focus returns semantic and syntactic information about the query that can be used by software to indicate where the data used in the query expression was located in the XBRL instance. The query-focus value is provided as a local variable. The rule focus must evaluate to a fact, facts, a concept or concepts. The inclusion of this component is optional. The processor will default to the first fact in the expression when processing the rule if a rule focus is not provided. The format of the results returned by query focus is processor dependent.

4.7.1.5.4 Query Suffix

The query-suffix is used to add a suffix to a query name. The query-suffix can be passed as a literal or a variable allowing variable names to be defined at run-time. A single query can generate query results with different query names depending on the input to the query. The query-suffix is concatenated to the end of the rule name using a period as a separator. The separator character can be changed using the query name separator declaration.

4.7.1.5.5 Output Attribute Value

An output attribute value can be added as a query part. The output attributes are defined as part of the output attribute declarations component of the query set. The values to be returned for each of these attributes is defined as part of the Query block. Output attributes must be returned as strings.

4.7.2 Example Queries

assert abc.0001 satisfied

exists({@ where $fact < 0 }#nonnegitem)

message 
"The value of {$nonnegitem.concept.name}  is equal to {$nonnegitem} for the period {$nonnegitem.period}. Please make the value positive.'"

severity error
query-focus $nonnegitem
query-suffix $nonnegitem.concept.name.local-name
myAttribute $nonnegitem.decimals.string

output abc.0002 

{@ where $fact < 0 }

4.8 Query Name Options

The query name prefix declaration is used where query names have a common prefix naming convention. A common prefix can be defined for all queries in the same file. The keyword query-name-prefix is used to set a standard prefix for all rules in the file where the declaration is defined. When query names are defined the prefix does not need to be included.

The query name prefix must appear preceding the queries in the file that it is intended to apply to. Multiple query name prefix can be defined in a file. Any queries after the declaration of the query name prefix will use the later prefix.

query-name-prefix my_rules

The Query name separator declaration is used to change the separator added by query-name-prefix`` andquery-suffix`. To change the separator from the default period to a colon the following declaration is made.

query-name-separator ":"

4.9 Compiling the Query Set

A query set zip file contains the compiled query set components in a zip file format. The query set zip file is not normative.

A query set zip file MUST conform to the [.ZIP File Format Specification][ZIP] (rpe:invalidArchiveFormat or tpe:invalidArchiveFormat).

5 Query Expressions

The query expression defines the logic of the query. The query expression must resolve to an iteration with a result or a non binding iteration. The query expression can comprise the following components:

• Parenthesis expression

• local variable definitions

• local variable references

• constant references

• custom function references

• for loops

• if-else statement

• operators

• filter expressions

• fact queries

• navigate queries

• collections

• property expressions

• function references

• literals

5.1 Literals

A literal represents strings, numbers , QNames and keywords.

5.1.1 String Literal

Strings are represented using either a single quote or double quote. Valid characters for a string is utf-8 characters.

5.1.1.1 Escape sequences in strings

The following escape characters are supported to allow values in strings to be escaped.

5.1.1.2 Substitute an expression in a string

Curly bracket substitution allows strings to be formatted with a query expression. Curly brackets are defined in the string and contain the query expression. Curly brackets can be used in a message but must be escaped.

5.1.2 Numeric Literals

Float and integer literals are supported. An integer can have an optional sign followed by numbers. No whitespace between the sign and the number is used to represent the sign. To define a number as a float it must include a decimal point. Floating numbers can include scientific notation.

Formatting of floats must include a decimal point and at least 1 digit. No digits have to follow the decimal point. No digits need to precede the decimal point. When using scientific notation there must be at least 1 digit to the right of the decimal point, which can be zero.

5.1.2.1 Inf literal

The inf literal represents a floating number. Inf is commonly used to represent infinite decimals of a fact. Inf can also be used to represent an infinite number. Inf is case-insensitive.

max(list(inf,1,2,3))

5.1.3 Period Type Literal

The period type can have literal values of either duration or instant. These are not defined as strings in an expression but are evaluated to a string. The evaluated string value is case sensitive. (lower case). The literal value is case insensitive.

5.1.4 Balance Type Literal

The balance type can have literal values of either credit or debit. These are not defined as strings in an expression but are evaluated to a string. The evaluated string value is case sensitive. (lower case). The literal value is case insensitive.

5.1.5 Severity Type Literal

The severity type can have literal values of either error, warning, ok or pass. These are not strings and are case insensitive.

5.1.6 Boolean Type Literal

The boolean type can have literal values of either true or false. These are not strings and are case insensitive.

5.1.7 Forever literal

Forever defines a duration object of unlimited length. It also represents periods that have no period. The literal is case-insensitive. Forever is not defined as a string in an expression and is evaluated as a forever duration. The literal value is case insensitive.

{@period = forever }

5.1.8 QName Type Literal

The Qname type must have literal values that represent QNames. These can be prefixed or un-prefixed.

5.1.9 Tagged Expressions

The # tag is used to associate the value before it with a tag immediately after it. The tag must be a simple name, and is case sensitive. The tag can then be used as a variable in any query output such as the query message using a $. The tag must not be referenced in the body of the query.

$ValueOfAssets = 20#Assets

30 + $Assets

5.1.10 Index Expressions

An index expression is used to retrieve data from a list or a dictionary. The index is represented using a square bracket notation.

For a list the square brackets indicate the location in the list. Items in a list start at 1. To retrieve data from a list the index expression must be an integer greater than or equal to 1. If the integer is larger than the length of the list or has a value less than 1 then an xbrlqe:IndexValueOutOfRange error must be produced.

list('a', 'b', 'c')[2]

For a dictionary the square brackets indicate the key of the key value pairs in a dictionary. If the index item does not appear in the dictionary then a none value is returned. The value of the key can be any type of object.

dict(list('0', 'zero string'), list(Assets, 'a QName'), list('a', 'a string'), list(0, 'zero number'))[Assets]  

If an index expression is used, on anything other than a list or dictionary then an xbrlqe:IndexNotOnListOrDictionary error must be produced.{#index-op-set-error}

5.1.11 Unary Characters

The unary characters of + or - can be used in front of a number to reverse the sign of a number or an expression.

5.1.12 Other Literals

The following additional literals can be used:

• skip
• none

These literals are not case sensitive.

5.2 Local Variable Definitions

All local variables are defined using a simple name.

A simple name is based on a NCname. A simple name extends the NCname by excluding the use of '.' in the name.

A local variable is used to store values within a single iteration of a query.

A local variable can be defined within a query expression.

A local variables type is not defined.

A local variable can represent any object in the Query Data Model

A local variable can represent a set, list or dictionary.

A local variable is assigned using the = operator or the in operator when used in a for loop.

A local variable is dynamically typed based on the value or object.

A local variable name is prefixed with a $ sign.

An expression defining a local variable should be concluded with a ;.

A local variable if reassigned within the scope of a processing iteration must follow the variable order of precedence.

Local variable values are not stored across processing iterations.

output abc.0003
$hello = 'hello';
$hello

5.3 Variable Order of Precedence

The same local variable can be defined multiple times.

Variables can be set in a number of ways and it may not be clear which value is assigned to variables with the same name. Variables can be defined by a constant, as an argument to a function, as part of a for loop or direct assignment to a variable.

A constant cannot have the same name as another constant but a local variable or function variable can. The constant has a lower order of precedence than a local variable or function variable if they have the same name.

$a = 10;
$a = 20;

function test($a)
$a = 30;
$a

constant $a = 40

$a = test(20) + $a

Constants can be referenced in a function but the value will be superseded by the argument to the function and this will be superseded by direct variable assignment in the function.

function test($a)
    $a

test(20) + $a 

function test($b)
    $a

test(20) + $a

5.3.1 Reserved Variables

The following variables are reserved for use when using a where clause expression:

• The $fact variable is used with the where clause used in a fact query.

• The $relationship variable is used with the where clause used in a navigate query.

• The $item variable is used with the where clause and returns component of a filter expression.

5.4 Iterative Statements (For Loops)

A for loop is used to iterate through a set or list of values. A for loop is defined using the keyword for.

There are a number of objects and properties that returns sets and lists that use the for loop to access the values. These include the following: - Network relationships property - Concept all-references property - Reference parts - Taxonomy networks property - Taxonomy concepts property Or any other set or list.

The for loop has the following structure:

for (variable in (set|list))
    Repetend

Where repetend is the expression to be repeated. For loops do not define how many times the loop should execute. The for loop will repeat until it gets to the end of the set or list. The set or list can be entered as a variable or as an expression. The for loop must loop through a set or a list. Any other object will cause the error xbrlqe:InvalidForLoopObject.

The variable declaration in the for loop is defined as a local variable and must follow the constraints of defining a local variable.

for ($c1 in taxonomy().concepts)
    $c1.name

The results of a for loop can be returned as another set by defining a variable.

$string_name_of_concepts = set(for ($c1 in taxonomy().concepts)
                    $c1.name.local-name)

A for loop can be used in the definition of a constant.

The same variable name must not be used where a for loop is contained within another for loop.

5.4.1 Assigning Local Variables in For Loops

A local variable can be declared by looping through a list or a set.

for $a in list('a', 'b')
        for $b in list('x', 'y')
            1
message '{$a} {$b}'

    'a x'
    'a y'
    'b x'
    'b y'

5.5 Conditional Statements (If-else statements)

An If-else statement is used to define conditional statements.

The if-else statement has the following form:

if ( condition ) statement1 else statement 2

The condition is a boolean expression. The condition MUST resolve to a boolean result. xbrlqe:InvalidObjectType

If the condition resolves to a value of none, the iteration is skipped.

Both the statement1 and statement 2 need to be included in the if-else statement.

The following example will return a true condition if the amendment flag in the default is set to true and there is no amendment description anywhere in the filing. If either of these conditions are false then the else statement returns false.

if ([@dei:AmendmentFlag] == true and count(list({covered @dei:AmendmentDescription})) == 0))
    true 
else
    false

If the else statement is to do nothing then the keyword none should be used or the iteration can be skipped using the skip keyword. The else statement must be included to indicate the end of the if condition.

The if-else statement condition can be encapsulated in parentheses.

An if statement can be included in a for loop.

for $x in set('a','b','c')
    if $x == 'a'
        false
    else
        true

5.6 Filtering Collections

The filter expression is used to filter a set or a list. A filter returns a set or a list depending on the collection type passed to it.

The filter expression MUST be passed a set or list to filter. If an object other than a set or list is used the xbrlqe:InvalidFilterExpression error is returned.

The railroad diagram in Figure 7 shows the components of a filter expression.

Figure 7: Filter Expression

filter $a

The filter expression uses a where clause to filter the list or set based on a boolean condition. The where clause uses the $item variable to represent the current value in the set or list.

filter set(1,2,3) where $item > 1

The values returned by the filter are specified by using the keyword returns. If returns is left off the values from the set not removed by a where clause are returned. The $item variable can be used in the returns expression.

filter $networkQname returns $item.local-name

The filter expression MUST return an error when an object other than a set of list is filtered. xbrlqe:InvalidFilterExpression - Filter expression can only be used on a 'set' or 'list'.

(filter $sub-periods returns "\t" + $item.period.string + "\t" + $item.string).join("\n");

Filter expressions can be nested inside other filter expressions. The $item variable remains local to the filter expression being referenced.

    filter (filter set(1,2,3,4) where $item > 1) where $item < 4

6 Fact Queries

To select a set of facts for evaluation a fact query is defined as part of the query. The railroad diagram below in Figure 8 shows the components of a fact query body.

The fact query will return XBRL facts defined as Facts, but excludes fact footnotes. These are facts with the concept name of note. These facts are queried using the footnote object properties.

FactQueryBody:

Figure 8: Fact Query Body

The fact query consists of components to control dimensional alignment (covered and covered dims), dictate the handling of nils, define dimension filters and define the where clause. These are combined to select facts from the XBRL instance.

The dimension filter is a component of the fact query body that defines the value of a dimension to match a fact. The dimension filter defines the dimension, an operator and a matching value called a dimension expression. The dimension filter is comprised of the components described in Figure 9 below.

Figure 9: Dimension Filter

All dimension filters in a fact query are prefixed with either the @ or @@ character. This indicates that the subsequent string is a dimension name.

The dimension filter expression has the following form:

@[Dimension Name] [operator] [Dimension Expression]

The operator and dimension expression are optional.

The operator must be one of the following: =, !=, in, not in

The dimension name is used to define the dimension of a fact to select on. The dimension name can be the dimension and/or the dimension properties. The dimension name values of the dimension filter must be one of the values described in Figure 10 below.

DimensionName:

Figure 10: Dimension Name

Dimension names must only appear once in a fact query.

A dimension name is comprised of the dimension and/or its dimension properties. The following represents different dimension names.

@concept.balance

@concept.period-type

If a dimension name is defined that is not in the above list or has a property that is not defined for the dimension name then the error xbrlqe:InvalidDimensionName is returned.

The Dimension Expression defines the object or QName of an explicit dimension or a value of a typed dimension. The expression must resolve to the type expected by the dimension name. When the in or not in operator is used the dimension expression can be a set or a list containing values with a type matching the dimension name. When the dimension dictionary filter is used the dimension expression must be a dictionary.

6.1 Concept Filters

A concept filter is a type of dimension filter that filters facts based on the concept core dimension. The concept filter is defined within a fact query.

The concept filter has a shortcut that allows the @ symbol to be provided with a QName immediately following the @. This shortcut eliminates the string '@concept ='

The concept filter @concept = Assets is the same as @Assets.

XBRL facts can be filtered based on the concept object and specific concept properties defined below. The table defines the allowable concept filters.

6.1.1 Selecting Collections

All of the selectors listed above can also be used with a set or a list using an in operator.

{@concept in list(Assets, Liabilities)}

6.1.2 Converse Selections

To select facts not matching a concept the != operator is used. To select all items not in a list the not in operator is used.

{@concept != Assets}

{@concept not in list(Assets, Liabilities)}

Only the = , != , in, and not in operators are used in a concept filter for a fact query .

Multiple concept filters can appear in a fact query. Multiple concept filters are treated as an and condition. The following query returns those facts where the concepts are an instant and have a debit balance attribute.

{@concept.period-type = instant @concept.balance = debit}

6.2 Period Filters

A period filter is a type of dimension filter that filters XBRL facts based on the period core dimension.

Multiple period filters can be defined in the fact query.

{@period.start = date('2026-01-01') @period.end = date('2027-12-31')}

Facts without a period dimension are filtered using the forever function or none.

{@period = forever} 

{@period = none} 

To take the period dimension out of alignment, the @period without an operator or value is used.

{@period} 

To take the period dimension out of alignment, without selecting forever periods the * value can be used as the filter value.

{@period = *} 

Allowable period values are a list, a set, a variable, a date, a duration, *, none, forever or no value.

6.3 Unit Filters

A unit filter is a dimension filter, that filters XBRL facts based on the unit core dimension.

The dimension name of the unit filter can either be the QName or the unit object.

The numerator and denominator properties of a unit cannot be used as apart of the <.

A unit object can be defined using the unit function. The expression below creates a USD/Shares unit object, that can be used to filter facts.

{@unit = unit(iso4217:USD, xbrli:shares)}

To select facts with a unit the * wildcard can be used as the filter value.

{@unit = *}

To select facts without a unit the keyword none can be used as the filter value.

{@unit = none}

To take units out of alignment the unit dimension filter can be used with no value. This will select all facts with units and without units.

{@unit}

Allowable unit values are a list, a set, a variable, a unit, *, none, or no value.

6.4 Entity Filters

An entity filter is a dimension filter, that filters XBRL facts based on the entity core dimension.

Facts can be filtered on the following components of the entity object.

An entity object can be defined using the entity function. The expression below creates a a entity object, that can be used to filter facts.

{@entity = entity('http://www.xxx.com','0000320193')}

To select facts with an entity the * wildcard can be used as the filter value.

{@entity = *}

To select facts without an entity the keyword none can be used.

{@entity = none}

To take entity out of alignment the entity dimension filter can be used with no value. This will select all facts with and without the entity dimension.

{@entity}

Allowable unit values are a list, a set, a variable, an entity object, *, none, or no value.

6.5 Taxonomy Defined Dimension Filters

There are two types of taxonomy defined dimension filters. The first is the taxonomy dimension filter. The taxonomy dimension filter is a dimension with a domain member or typed value paired with a dimension, expressed as dimension = member.. The second is a dictionary of dimensions with a domain member or typed value pairs. This is called a dimension dictionary filter.

6.5.1 Taxonomy Dimension Filter

To return all facts using a specific dimension the wildcard * is used with the dimension name. This does not return facts with no dimensions as the default dimension member is not included in the wildcard.

The wildcard returns facts that have additional dimensions in addition to the defined dimension name.

{@dei:LegalEntityAxis = *}

{@dei:LegalEntityAxis = none}

{@dei:LegalEntityAxis != *}

To take the dimension out of alignment the dimension filter can be used with no value. This will select all facts with the taxonomy defined dimension as well as facts with the default dimension member. This will exclude facts with other taxonomy defined dimensions.

{@dei:LegalEntityAxis}

6.5.2 Dimension Dictionary filter

The dimension dictionary filter has a dimension expression that is a dictionary. The keys of the dictionary are the taxonomy defined dimensions, the values are the taxonomy defined domain members or typed members defined in the instance.

The dictionary can contain typed and explicit dimensions.

The dictionary value associated with the key must be a domain member or a typed value. The value of the typed dimension member must match the type of the typed dimension.

The in and not in comparison operators must not be used with dimension dictionary filter and a dimension expression in a list or a set.

    $Assets = {@Assets}.dimensions;
    {@concept = Liabilities @dimensions = $Assets}

    $Assets = {@Assets}.dimensions
{@concept = Liabilities @@dimensions = $Assets}

The dimensions filter only applies to taxonomy defined dimensions.

The dimensions filter does not include the period, concept, entity or unit dimensions.

The dimensions filter is overridden by any explicit taxonomy defined dimension filters defined in the fact query expression that conflict with the values defined in the dimensions filter.

Only one taxonomy defined dimension filter can be defined in a fact query.

    {@concept = Liabilities @dimensions}

    {covered-dims @concept = Liabilities}

The taxonomy defined filter can be defined with a wildcard to select facts that have taxonomy defined dimensions. This excludes any facts in the default dimension.

    {@concept = Liabilities @dimensions=*}

The taxonomy defined filter can be defined with a none value to select those facts that have no taxonomy defined dimensions. This will include only those facts in the default.

    {@concept = Liabilities @dimensions=none}

    [@concept = Liabilities]

Specific taxonomy defined dimensions can be used with the dimensions filter. If they conflict, the named dimension takes precedence. If the dimension filter equals none, an additional filter is not considered a conflict.

    {@concept = Liabilities @dimensions=none @dei:LegalEntityAxis = *}

    [@concept = Liabilities @dei:LegalEntityAxis = *]

6.6 Cube Filter

A cube filter selects facts that appear in a hypercube. The hypercube can be defined using the cube function by passing the hypercube concept and drs role as parameters to identify the hypercube.

Cubes can also be specified by the hypercube concept name, the drs role of the cube or both. This selection below returns all facts associated with cubes named StatementTable.

{@cube.name = StatementTable}

Facts in a cube can also be specified by defining the drs role.

{@cube.drs-role = BalanceSheet}

This returns facts in the cubes in drs role BalanceSheet.

Using @cube has no effect on the alignment of the returned facts. @cube is only used to filter facts for a fact query .

To return all facts associated with any cube use one of the following:

{@cube != none}

{@cube = *}

To return all facts not in a cube use the equal none expression.

{@cube = none}

6.7 Instance Filter

An instance filter selects facts that appear in a named instance object. The instance can be defined using the instance function by passing the instance uri. An XBRL instance is defined as an XBRL report.

The instance filter filters on the instance of the current filing being evaluated. The instance filter> allows multiple instances to be processed simultaneously. If no instance filter is defined then the instance provided to the processor is evaluated.

$wsfs = instance('https://www.sec.gov/edgar/wsfs-20211231.htm')

Fact query selection defaults to the instance document passed to the processor. To select facts from an alternative instance document the fact query selection criteria must reference the instance object.

{@instance = $wsfs}

Sets of facts may be returned from multiple instances using a list of instance objects.

{@instance in list($wsfs, $appl)}

The default instance may be specified using instance()

{@instance = instance()}

{@instance in list(instance(),$wsfs,$appl)}

The wildcard and not in selectors are invalid when filtering on the instance dimension.

{@instance not in list(instance(),$wsfs,$appl)}

{@instance = *}

6.8 Language Filter

A language filter is a dimension filter, that filters XBRL facts based on the language core dimension.

The dimension name of the language filter must be a BCP47 language code.

{@language = 'fr'}

To select facts with a language dimension the * wildcard can be used as the filter value.

{@language = *}

To select facts without a language the keyword none can be used as the filter value.

{@language = none}

To take language out of alignment the language dimension filter can be used with no value. This will select all facts with and without a language dimension defined.

{@language}

Allowable language values are a list, a set, a variable, a unit, *, none, or no value.

6.9 Square and Curly Bracket Control

Facts returned can be controlled using [] characters and {} characters in the fact query.

A curly bracket notation is used to include facts with taxonomy defined dimensions in the fact query.

A square bracket notation can be used to return those facts without taxonomy defined dimensions. The square bracket and curly bracket notation is used independently.

Figure 11 below shows that the fact query body can be included in square, curly or no brackets.

Figure 11: Fact Query

A fact query with no brackets is the equivalent to curly brackets.

    [@] or []

    [@concept = Assets]

    [@concept = Assets @LegalEntityAxis = *]

    {@concept = Assets @LegalEntityAxis = *}

The @ character with no dimension name defined will return all facts. The @ can stand by itself as shown in the following examples. Curly brackets are optional.

{@}  or  {}  or  @ 

[@]  or  []  

6.10 Filter Control

6.10.1 Dimensional Alignment Control

All facts returned in a fact query have an associated dimensional alignment. Dimensional alignment is a dictionary with each dimension representing a key and each dimensional value representing a dictionary value. When performing operations on facts, the operations are performed between facts with the same dimensional alignment, this is called binding. All facts with the same dimensional alignment must bind. Each calculation defined will be performed for each dimensional alignment returned from the fact query. Each dimensional alignment will generate an iteration. Facts with no dimensional alignment dictionary will bind with facts in every iteration.

Literal values such as strings and integers defined in a rule have no dimensional alignment.

Dimensional alignment can only be associated with a fact.

Dimensional alignment may be removed from facts returned in a fact query using the covered keyword.

6.10.1.1 Covered Filters

Facts returned using a fact query may have the dimensional alignment removed by using the keyword covered in the fact query. The covered keyword will set the alignment to none. Facts with a dimensional alignment of none will bind with any other fact irrespective of their dimensional alignment. The keyword covered should be used before the dimension filters in the fact query expression.

{covered @concept = Assets}

6.10.1.2 Using the At Sign (@) to Cover Specific Dimensions

The @@ syntax acts as a filter but also maintains the dimensional alignment of the fact. The @ syntax removes the dimension from the dimensional alignment of the fact.

Where multiple filters are used for the same dimension but have different alignment controls the processor will default to maintaining dimensional alignment and the double @ takes precedence.

@concept.balance = debit @@concept.period-type = instant

@@concept.balance = debit @@concept.period-type = instant

{@concept @entity @period @unit @dimensions}

The fact derived using @ on every dimension will only bind with other facts that have no dimensional alignment, but not with facts that have dimensional alignment.

The double @@ forces dimensional alignment on a filter and can be used to uncover a specific dimension when it is covered using the keyword covered or covered-dims. It can also be used to cover a specific taxonomy defined dimension when the @dimension filter is used.

list({covered @RevenueFromContractWithCustomerIncludingAssessedTax @@srt:ProductOrServiceAxis = *})

6.10.1.3 Using covered-dims to cover dimensions

Taxonomy defined dimensions can be covered using the keyword covered-dims.

The keyword covered-dims removes taxonomy defined dimension alignment from facts returned by a fact query. Facts that are covered with covered-dims maintain their core dimensional alignment. The keyword covered-dims should be used before the dimension filters in the fact query expression.

    {covered-dims @concept = Assets}

6.11 Handling Nils

Facts can be defined in an instance with a nil value. Nils have no numeric value or decimal value.

The processor by default includes nil values in the fact query results. A processor can be instructed to exclude nil values from the fact query when processing is initiated. A fact query can be defined to exclude nil values by using the keyword nonils. The keyword nonils removes facts with a nil value from the fact query .

The keyword nils can be set to include nil values in the fact query. This is used if the processor is set to exclude nils A processor can be instructed to exclude nils from fact queries. The nils keyword will override the processor instructions.

A fact with a nil value will fallback to a value of zero for numeric facts if the nildefault keyword is used in the fact query.

The value returned for nil when the nildefault keyword is used differs depending on the type of the concept. If the concept is a numeric type then a value of zero is returned. If the concept is non numeric then an empty string is returned.

Any other type will return as nil.

6.12 Defining a Dimension Alias

A dimension filter can be referenced by defining an alias in the dimension filter. The alias name is defined as a local variable and is defined using the as keyword. The as keyword allows a dimension to be defined as an alias that can be used to reference a dimension filter in a where clause.

The alias name must be a simple name.

An alias makes it easier to handle expressions in the where clause as the names are shorter and easier to read.

{@concept = Assets @dei:LegalEntityAxis=* as $lea }

{@concept = Assets  @dei:LegalEntityAxis =* as $lea where $fact > 0 and $lea == ParentCompanyMember }

6.13 Where Clause

The where clause allows filtering of returned facts based on properties of the fact. The where clause can filter on all properties of the fact and has no impact on the dimensional alignment of the fact.

To evaluate fact query results with the where clause, use the $fact variable to refer to the attributes of the filtered fact.

The $fact variable is a built in alias representing the fact object returned from the fact query. All the properties of a fact object can be returned from $fact and operated on as part of the where clause.

The where clause must be contained within the fact query .

{@ where $fact < 0} 

{@unit ! = xbrli:pure where $fact < 0 and $fact.decimals == -6}

6.14 Nested Filters

Nested filters allow operations to be defined at different levels of dimensional alignment within an iteration.

Under normal processing all operators are performed within a single dimensional alignment. This alignment normally represents a single iteration. With nested filters operations with multiple dimensional alignment can be performed in a single iteration. This means results with one dimensional alignment can be compared to results with a different dimensional alignment in the same iteration.

Nested filters group fact queries within other fact queries. Dimensional alignment is performed at the lowest level. The results then roll up to the next level of the fact query.

{@actualMonthlyPayment} - {@actualMonthlyReimbursement}

{@contractedMonthlyPayment} - {@actualMonthlyPayment} - {@actualMonthlyReimbursement}

{@contractedMonthlyPayment @period} -  {@actualMonthlyPayment} - {@actualMonthlyReimbursement}

{{@actualMonthlyPayment} - {@actualMonthlyReimbursement}}

{@period {@actualMonthlyPayment} - {@actualMonthlyReimbursement}}

{@contractedMonthlyPayment @period}  - {@period {@actualMonthlyPayment} - {@actualMonthlyReimbursement}}

6.15 Fallback values

If a fact query does not produce a result it will skip the iteration. If a fact query has a dependence on another fact query value in the expression that does return a result the empty fact query will fallback to a value of zero. The default fall back value of zero can be be changed using the first-value function (see Section 11.10.3). The first-value function allows a sequence of fallback values to be defined.

{@concept = Assets} - {@concept = Liabilities}

{@concept = Assets} - first-value({@concept = Liabilities},{@concept = LiabilitiesCurrent},100)

7 Taxonomy Navigation

Taxonomy Navigation is used to query relationships in a taxonomy.

The navigate query is used to instruct the processor to navigate the relationships in a taxonomy and return the relationship objects encountered as it follows each path as either a set, list or dictionary.

The navigate query has the following components:

• arcrole selection

• direction and level of navigation

• root relationship

• start point of navigation

• end point of navigation

• conditional stopping of navigation

• selecting roles to navigate

• selecting taxonomies to navigate

• where clause filtering of results

• result options

• collection type returned

The railroad diagram below describes how the navigate query components are sequenced and the permitted combinations.

Figure 12: Navigate query

The navigate function will navigate the trees until it reaches the end of the tree or encounters a concept it has already encountered. (A cycle)

If a cycle occurs the cycles attribute is flagged indicating the that navigation encountered a cycle. The processor will then stop navigation at that point. The relationship where the repeated concept is the target will be returned.

7.1 Selecting Arcroles to Navigate

Navigation can be limited to specified navigation arcroles. To define the arcrole to navigate either the full uri or the last path component of the arcrole uri can be used.

navigate parent-child descendants

navigate parent-child descendants

navigate 'http://www.xbrl.org/2003/arcrole/parent-child' descendants

If the last path component of the uri is not unique within the taxonomy then all uri's with the same last path component are used as the arcrole parameter.

The arcrole navigation parameter can be a uri, a string, an arc-role object, a set or a list.

Multiple arcroles can be defined in a set or a list.

The arcrole parameter is optional in the navigate query.

7.2 Direction of Navigation

The navigation direction parameter indicates the direction taken when traversing arcroles defined in a taxonomy hierarchy.

Navigation direction must have one of the values described in Figure 13 below.

Figure 13: Navigation Direction

The descendants navigation direction instructs the processor to continue down all branches of the arcrole hierarchy from the starting concept until there are no more child concepts, or a stop condition is reached.

The children navigation direction instructs the processor to traverse down 1 navigation step of the arcrole hierarchy from the starting concept.

The ancestors navigation direction instructs the processor to continue up all branches of the arcrole hierarchy from the starting concept until the root concept is reached, or a stop condition is reached.

The parents navigation direction instructs the processor to traverse up 1 navigation step of the arcrole hierarchy from the starting concept.

The siblings navigation direction instructs the processor to select sibling relationships of the starting concept in the arcrole hierarchy. The sibling relationship returned is the relationship where the sibling is the target.

The previous-siblings navigation direction instructs the processor to select sibling relationships of the parent concept of the starting concept in the arcrole hierarchy.

The next-siblings navigation direction instructs the processor to select sibling relationships of the children concept of the starting concept in the arcrole hierarchy.

The source and target attributes of a relationship are not impacted by the navigation direction.

The number of relationships to navigate in a given direction are called navigation steps.

When the direction is defined as descendants or ancestors, the number of navigation steps can be specified after the navigation direction as an integer. If no navigation steps are defined then the navigation function continues down all branches until it reaches a leaf node.

navigate parent-child descendants 1

navigate parent-child children

The navigation direction parameter must be defined when using a navigate query.

By default, navigation returns the target concepts of the relationships exclusive of the starting concepts.

7.3 Root Relationship

The include start keyword is used to return a synthetic relationship defining the root or starting concept as a relationship target concept.

To include root concepts, the include start keyword is defined after the navigation direction parameter.

navigate parent-child descendants include start

navigate parent-child descendants 2  include start

navigate ancestors

7.4 Start Point of Navigation

By default, navigation starts at the roots of an extended link role. The from concept instructs the navigate query to start navigation at a particular concept in a hierarchy. The keyword from preceding the concept object or concept QName indicates the starting concept. Concept properties cannot be used as a parameter of the from concept.

    navigate parent-child descendants 2 from Assets

7.5 End Point of Navigation

To define an end point of a navigation traversal a to concept is specified. The to concept instructs the navigate query to stop navigation at a particular concept in the hierarchy. The keyword to preceding the concept object or concept QName indicates the ending concept. Concept properties cannot be used as a parameter of the to concept.

If to concept is provided with the keyword 'to' then the relationships between the starting concept (or root concept) to the 'to concept will be returned.

The relationship where the target concept equals the to concept will be returned.

If a hierarchy is navigated and the 'to' concept is never reached then no relationships will be returned.

navigate parent-child descendants from Assets to OtherAssetsCurrent

7.6 Conditional Stopping of Navigation

The stop when keyword is used to stop navigation of a hierarchy when properties of the relationship equal a value defined in the navigate query. A navigation of a hierarchy is stopped when the keyword stop when is followed by a boolean expression that resolves to true.

The boolean expression evaluates if an attribute of the relationship matches a condition. This enables navigating a hierarchy and stopping navigation down a given branch when a condition is met on the relationship. The stop when keyword differs from the 'to' keyword in that all relationships that evaluate to false will be returned. Using the 'to' concept will only return relationships if the 'to' concept actually exists in the hierarchy.

The relationship that evaluates to true when using stop when will be returned as part of the result.

navigate parent-child descendants from IncomeStatementAbstract stop when $relationship.target.name == GrossProfit

7.7 Selecting Roles to Navigate

Relationships in a specific extended link role can be navigated by specifying the navigation role. The keyword role is followed by the role object, the role uri or the last component of the uri path. When the last path component of the role is used, the last path component must be unique within the taxonomy.

navigate parent-child descendants role 'http://www.abc.com/role/ConsolidatedBalanceSheets'

navigate parent-child descendants role ConsolidatedBalanceSheets

The last path component of the role must not use quotes.

7.8 Selecting a Linkbase

To restrict navigation to a specific linkbase the linkbase navigation keyword is used. The linkbase keyword is followed by a linkbase link name. The link name must be defined as a QName. If a QName is not provided the xbrlqe:InvalidLinkbaseNavigationObject error is returned.

navigate descendants from Assets linkbase link:calculationLink taxonomy $US-GAAP-2020

7.9 Selecting Taxonomies to Navigate

By default, navigation occurs in the taxonomy of the instance document. The navigation taxonomy keyword is used to navigate a specific taxonomy other than the default taxonomy associated with the instance. The keyword taxonomy is used followed by the taxonomy object.

navigate parent-child descendants from Assets taxonomy taxonomy('http://xbrl.fasb.org/us-gaap/2016/entire/us-gaap-entryPoint-std-2016-01-31.xsd')

navigate parent-child descendants from Assets taxonomy $myInstance.taxonomy

7.10 Where Clause selection on Navigation Results

Navigation returns relationship information based on the relationship found during the traversal. A navigation where clause expression can be used to filter the relationships returned. For each relationship object, the where expression is evaluated.

If the 'where' expression for a relationship evaluates to true, the relationship is included in the result.

A special variable $relationship is used in the relationship where clause expression to refer to the properties of the relationship being filtered.

navigate parent-child descendants from Assets where not $relationship.target.is-abstract

7.11 Returning Navigation Results

The default result of navigation returns a set of target concepts from the relationships matching the criteria of the navigation query.

The returns keyword is used to specify the return components of the relationship to return.

navigate parent-child descendants from AssetsAbstract returns (source)

set(AssetsAbstract,CurrentAssets,NoncurrentAssetsNoncurrent,Land,Cash,Assets)

The following return components may be specified:

The source keyword instructs the processor to return the relationship source concept of the relationship.

The source-name keyword instructs the processor to return QName of the source concept of the relationship.

The target keyword instructs the processor to return the target concept of the relationship.

The target-name keyword instructs the processor to return the QName of the target concept of the relationship.

The order keyword instructs the processor to return the relationship order property of the relationship.

The weight keyword instructs the processor to return the relationship weight property of the relationship.

The preferred-label keyword instructs the processor to return the label object for the label indicated by the preferred label for the target concept. This includes uri, description, lang and text properties.

The preferred-label-role keyword instructs the processor to return the role object for the preferred label. This includes the uri, description and used on properties. The relationship keyword instructs the processor to return the relationship object.

The role keyword instructs the processor to return the extended link role of the network where the relationship is defined.

The role-description keyword instructs the processor to return the description of the role of the network.

The arcrole keyword instructs the processor to return the relationship arcrole of the network the relationship is in.

The arcrole-description keyword instructs the processor to return the relationship arcrole-description property of the relationship.

The arcrole-cycles-allowed keyword instructs the processor to return the cycles allowed attribute of the arcrole definition. One of: 'any', 'undirected', 'none'.

The link-name keyword instructs the processor to return the QName of the extended link element.

The arc-name keyword instructs the processor to return the QName of the arc element.

The network keyword instructs the processor to return the relationship network property the relationship is in.

The cycle keyword instructs the processor to return the cycles attribute, a boolean indicator if the relationship starts a cycle in the navigation.

The navigation-order keyword instructs the processor to return the calculated sibling order of the relationship target. This is not the order on the relationship but is calculated during the navigation.

The navigation-depth keyword instructs the processor to return the depth of the relationship target concept from the starting concept.

The result-order keyword instructs the processor to return the order of the result within the full result list.

Defining a QName as an arc attribute instructs the processor to return the value of the arc attribute. The actual QName of the arc attribute is used in the returns statement.

returns (source-name, target-name, ex:specialAttribute)

The dimension-type keyword instructs the processor to return the purpose of the target concept in dimensional navigation. See Dimensional Navigation.

The dimension-sub-type keyword instructs the processor to return the more specific purpose of the target concept in dimensional navigation. See Dimensional Navigation.

The drs-role keyword instructs the processor to return the initial role of the dimensional relationship set. See Dimensional Navigation.

Multiple return components may be returned by composing them in parenthesis.

navigate parent-child descendants from Assets returns (target, preferred-label)

The include start keyword creates an extra relationship result for each starting concept returned as the target. This affects the way the following return components are returned:

7.11.1 Returning by Network

Unless otherwise specified the result of navigation is a flat list of items. The by network navigation keyword groups the relationships by the networks that they are located in. The by network keyword returns a dictionary where each network is the key and the value of associated with the key is a list of the specified return items.

The by network keyword is used to separate relationships by the extended link role. Figure 14 below graphically shows two extended link roles with relationships.

Figure 14: By Network

    navigate parent-child descendants returns by network

7.11.2 Returns as Set or List

The return components from a navigate can be specified to indicate if they are returned as a set or a list. By default return components are returned as a list.

The returns-set keyword will return the values from evaluating the navigate function as a set rather than as a list.

The returns-list keyword will return the values from evaluating the navigate function as a list.

navigate summation-item descendants from  PropertyPlantAndEquipmentGross returns set (target-name)

7.11.3 Returning Paths

Default navigation returns a list of results. Alternatively, the results can be organized by the path of the navigation. The navigation paths keyword returns all of the relationships identified while traversing individual paths in a tree hierarchy. The path keyword results in path lists contained within an outer list. The outer list contains a list for each path of traversal. The inner list contains each relationship in the order of the traversal.

Figure 15 below graphically shows a hierarchy that has 4 paths if the starting node is 'A'.

Figure 15: Paths

The diagram has 4 paths that are defined in a list of lists when using the paths keyword.

    navigate parent-child descendants include start from A returns paths

list(
    list(A,B,D),
    list(A,B,E),
    list(A,C,F),
    list(A,C,G)
    )

    navigate parent-child descendants include start from A returns paths (source, target, order)

list(
    list(
        list(None, A, None),
        list(A,B,1),
        list(B,D,1)
        )
    list(
        list(None, A, None),
        list(A,B,1),
        list(B,E,2))
    list(
        list(None, A, None)
        ,list(A,C,2)
        ,list(C,F,1)
        )
    list(
        list(None, A, None),
        list(A,C,2),
        list(C,G,2)
        )
    )

    product(navigate summation-item descendants include start from ProfitLoss to Revenues returns paths (weight))

7.12 Return as Dictionary

The default result value is a list. When multiple return components are returned, it can be more useful to have the result returned as a dictionary. The as dictionary keyword is used to structure returned results from a navigate function as key-value pairs where the keys are the names of the return components.

    navigate parent-child descendants from Assets returns (source, target, role) as dictionary

7.13 Dimensional Navigation

The dimensional relationship sets (DRS) comprise relationships from multiple arc roles and extended link roles to form a model of the cubes defined in the taxonomy.

Dimensional navigation is used to navigate the DRS. Figure 16 shows how the dimension arc roles are composed into the DRS

Figure 16: Dimensional Navigation

Dimensional navigation traverses the multiple arc roles that make up a DRS. For the purpose of dimensional navigation, the hypercube is the root of the structure. This differs from the standard XBRL dimension model which treats the primary item as the root of the structure. Note that the all arc role in the XBRL dimension model is from the primary item to the hypercube. In dimensional navigation, this is flipped and the all arc role is treated as from the hypercube to the primary item.

The dimensions keyword is used to indicate that the navigate expression should use dimensional navigation.

navigate dimensions descendants from dei:LegalEntityAxis

Navigation is constrained to a single cube by using the cube keyword.

The cube keyword must only be used with dimensional navigation.

navigate dimensions descendants from dei:LegalEntityAxis cube us-gaap:StatementTable

Arc roles can be used to limit the results of dimensional navigation to only those relationships with the specified arc role.

navigate dimensions dimension-domain descendants from us-gaap:StatementTable

7.13.1 Pseudo Arcroles

In addition to the arc roles that are used to define dimensional relationships, these additional pseudo arc roles can be used.

The hypercube-primary arcrole will return relationships between a cube concept and the primary item concept. This arc is similar to the 'all' arc role to define a relationship between a primary item and a cube, but in the opposite direction.

The dimension-member arcrole will return domain-member relationships that stem from a dimension concept via a dimension-domain relationship. The arcrole only includes member concepts that are members of a dimension (not a primary item).

The primary-member arcrole will return domain-member relationships that stem from the primary concept of a cube. The arcrole only includes member concepts that are members of a primary item (not a dimension).

Any of the standard dimension arc roles, except for all and not-all, may be used for dimensional navigation. In dimensional navigation, the domain-member arc role applies to both members of a dimension and the members of the primary item.

When using an arcrole or pseudo arcrole the navigation will still traverse the DRS from the starting concepts regardless of the arc role specified. Only the relationships from the specified arcrole will be returned.

navigate dimensions dimension-member descendants from us-gaap:StatementTable

7.13.2 DRS Role

Dimensional navigation can traverse more than one extended link role. The original role used on the 'all' relationships between the primary item concept and the hypercube concept is the drs-role. This remains the same when navigating a DRS even though the extended link role may change.

To limit dimensional navigation to only one drs role the drs-role keyword can be used.

navigate dimensions descendants from us-gaap:StatementTable drs-role BalanceSheet

The drs-role can be specified using the uri of the role in quotes or the last path component. When the last path component of the drs-role is used, the last path component must be unique within the taxonomy, otherwise the error xbrlqe:AmbiguousExtendedLinkRoleURI is returned.

7.13.3 Navigating Cubes

To restrict Dimensional navigation to a specific hypercube concept the cube navigation keyword is used. The cube keyword is followed by either a hypercube concept, QName or cube object. If any other object is provided the xbrlqe:InvalidCubeNavigationObject error is returned.

The cube keyword can only be used when using Dimensional navigation. If the navigation keyword dimensions is not provided the xbrlqe:CubeCanOnlyBeUsedInDimensionalNavigation error is returned.

navigate dimensions children cube ScheduleOfFairValueOffBalanceSheetRisksTable taxonomy $US-GAAP-2020

7.13.4 Dimension Return Components

In addition to the non-dimensional navigation return components, dimensional navigation can use the following return components:

The dimension-type return component identifies the dimensional purpose of the target concept. The values are: * hypercube * primary-member * dimension * dimension-member

The dimension-sub-type return component identifies the specific dimensional purpose for dimensions and members. The values are:

1. For dimension-type = dimension

   • explicit - explicit dimension concepts
   • typed - typed dimension concepts

2. For dimension-type = dimension-member

   • default - the default member of an explicit dimension

3. For dimension-type = primary-member

   • primary - the primary item

The drs-role return component identifies the network og the hypercube. The extended link role of the primary to hypercube relationship (all relationship). In a dimensional relationship set, the role can be different for different relationships. The drs-role is constant for relationships that make up the dimensional model of a cube.

The useable return component identifies the value of the usable attribute. If there is no usable attribute it defaults to true.

7.13.5 Navigating Across Networks

In some cases it is necessary to navigate between concepts irrespective of the extended link role they are contained within. The navigation across networks keyword returns relationships irrespective of the extended link role they are associated with. Using this keyword will return all relationship associated with a concept. If traversing multiple paths, the paths are combined across all extended link roles into one network.

Figure 17 shows two networks in different roles with the same arcrole.

Figure 17: Separate Networks

The across networks command will run the navigate as if the network is as follows:

Figure 18 shows a single network with the same arcrole.

Figure 18: Combined Networks

8 Processing Model

8.1 Iterations

The processing model is an iterative model. The query expressions defined are processed iteratively. The number of times a query runs is dependent on the number of facts selected using a fact query and/or iterations created by processing for loops defined in a query.

A query starts execution with a single iteration called a start iteration. Queries can add additional iterations using a for loop or by running a fact query which selects any number of facts. The number of iterations for a fact is dependent on the number of contexts.

Once a for loop starts the start iteration becomes the first iteration of the for loop.

output abc
'hello'

output abc
for $x in list()
    'hello'

output abc
for $x in list(1)
    'hello'

When using a fact query the start iteration becomes the iteration of the first fact.

output abc
if exists({@concept = Assets})
    'hello'
else
    skip

The existence of a fact must be evaluated within an iteration. Every fact based iteration is run within a dimensional alignment.

Dimensional alignment is defined by the dimensions that a fact exists in. Every fact must have dimensional alignment. When calculations are performed between facts, calculations will only be performed between facts that are in a common dimensional alignment. The number of iterations performed will mirror the number ofdimensional alignments of the facts selected.

The start iteration has no dimensional alignment. To test if a fact exists within an instance document, it must be evaluated in the start iteration. This means the fact has to be taken out of dimensional alignment.

output abc
if exists({covered @concept = Assets})
    'Assets Exists'
else
    'Assets is absent'

The covered keyword removes the dimensional alignment from a fact. This allows the expression to be evaluated in the start iteration. If facts for assets do exist, the query will replace the start iteration with the fact iteration.

output hello 

'hello ' + string([@Assets])

As each iteration completes an output is generated and the next iteration starts. Iterations can be collected in a set or list and output as a single output. However the fact alignment of the iterations has to be the same, (if not a for loop) as each list or set also has an associated dimensional alignment.

8.1.1 Skipping an Iteration

A skip keyword is used to skip an iteration generated by an aligned fact or by a for loop. When the processor encounters a skip keyword the current iteration is dropped. If within an iteration the skip key word is encountered the iteration will be dropped and the processor will move onto the next iteration.

A skip can be used to skip a loop in a for loop.

Iterations are also be skipped in the case of specific evaluations on none values. These are:

• none and true

• none and none

• none or false

• none or none

• Integer / none

• Integer * none

8.1.2 Aggregation of iterations

The results of iterations can be aggregated together by including the results of iterations in a collection such as a set or a list.

output agg
list(@concept in list(CurrentAssets,NoncurrentAssets))

Lists and sets can aggregate iterations together. Aggregation operators can also aggregate iterations.

8.2 Alignment

Operations are performed on facts in iterations. Operations will only be performed on facts with the same alignment. Given facts have a unique dimensional alignment within an XBRL instance, an expression must remove alignments between facts to perform operations on them. When fact queries define a dimension filter that dimension is removed from dimensional alignment.

output facts 

{ where $fact > 0}

output facts 

$assets = {@concept = CurrentAssets} + {@concept = NoncurrentAssets}
$assets

output SumOfFacts 

$SumOfFacts = sum(list({@concept = * @period = * @SibAxis = *}))
$SumOfFacts

output SumOfFacts 

$SumOfFacts = sum(list({@concept = *  @SibAxis = *}))
$SumOfFacts

8.2.1 Alignment of Dimensions

Alignment represents the dimensions associated with a fact. This includes the non-taxonomy dimensions of concept, period, entity and unit as well as the taxonomy dimensions.

By specifying a dimension in a query it can be selected. The selection of the fact by a given dimension drops that alignment for the fact.

output AssetsAlignment 

{ @concept = Assets}

output AssetsAlignment 

{ @concept = Assets @unit=unit(iso4217:NZD)}

output AssetsAlignment 

{ @concept = Assets @unit=unit(iso4217:NZD) @period = date('2085-12-31')}

output AssetsAlignment 

{ @concept = Assets @unit=unit(iso4217:NZD) @period = date('2085-12-31') @Geography = AU}

output AssetsAlignment 

{ covered @concept = Assets}

By adding dimensional qualifiers to a selection of facts will control the alignment of facts. Alignment has limited impact on an individual fact, but becomes important when aligning facts with different dimensions in a query.

8.2.2 Fact Properties Post Operation

The result of adding, subtracting or applying another operator to a fact will result in a value that has none of the properties of the originating facts. The resulting value will not have any built in dimensions, taxonomy defined dimensions or decimals. The resulting value will be a scaler value. The scaler value will still have alignment of the facts it was generated from, and will be able to participate in operations with other scalers or facts with the same alignment.

The query language cannot generate new facts from existing facts in the instance that are available to the processor as a fact object.

8.2.3 Scalers with no alignment

Scalers with no alignment such as an integer or decimal defined in an expression can participate in operations with values that do have alignment. A value with no alignment is distinct from a value with empty alignment. Empty alignment occurs when dimension filters remove all the dimensions in the dimensional alignment.

8.3 Processing nil and none Values

A nil value occurs when a fact in the XBRL instance has a value of nil. The fact query allows the binding of nil bindings to be defined.

A none value represents an empty result from evaluating an expression. A none value for a fact object is returned when a nil fact value is encountered in an instance document.

sum(set())

dict(list(Assets, 'a qname'), list('a', 'a string'))['kitten']  

8.3.1 Nil Operations

When performing operations a none value defaults to the following when operated against the defined type.

The table list how nil values must be processed with different operators.

The table lists how nil values must be processed with different functions

Properties of a none value will return a value of none. The property is-fact (see Section 11.8.2) will return a value of false.

When an If-else statement has a condition that resolves to none then the expression resolves to an unbound condition and the iteration is skipped.

See functions section for how each function handles none values.

8.4 Object Model

The processing model is based on the object model. The processing model operates on the object model independently of the syntax of the instance or taxonomy documents.

9 Collections

A collection is used to refer to sets, lists and dictionaries. XBRL query supports sets, lists and dictionaries and operations between them.

• A list is a group of ordered items.

• A set is a group of unique items.

• A dictionary is a group of items with key names that must be unique.

9.1 Sets

A set is an unordered collection that cannot include duplicates. Sets can be used to remove duplicates and to test for membership in a set. Sets allow the use of mathematical operations like union, intersection, difference, and symmetric difference. Sets cannot be sorted.

Any duplicate values assigned to a set are removed. If a covered fact query is used to derive a set, any facts with the same value are removed from the set, even if other attributes of the fact are different. Facts in a collection should generally use a list as an aggregator and not a set. A set is defined using the set() function.

9.1.1 Set Specific Operations

9.1.1.1 union(set to union) : set

The union operation will produce a union of two sets, eliminating any duplicate items contained in either set.

The union of two sets can also be performed using the + operator.

A union performed using the + operator must only be performed between two sets. A union performed using the + operator between two incompatible collections results in a xbrlqe:IncompatibleOperator error.

A union performed using the union property can use a list as an argument. The processor will cast the list in the argument to a set. The union property must only be used as a property of a set. xbrlqe:InvalidObjectProperty

$A = set(a,b,c);
$B = set(c,d,e);
$A + $B

$A = set(a,b,c);
$B = set(c,d,e);
$A.union($B)

set(a,b,c,d,e)

When the union is used as a property of none a value of none will be returned.

none.union(set(1,2,3))

When the + operator is used to union none with a set the value of the set will be returned. The + operator has special handling for missing data. If one side of the + operator is none then the value of the set is returned.

none + set(1,2,3)

9.1.1.2 intersect(set to intersect) : set

The intersect operation will produce an intersection of two sets, eliminating any duplicate items contained in either set.

The intersection of two sets can also be performed using the & or intersect operator.

A intersection performed using the & or intersect operator must only be performed between two sets. An intersection performed using the & or intersect operator between two incompatible collections results in a xbrlqe:IncompatibleOperator error.

An intersection performed using the intersect property will cast the list in the argument to a set. The intersect property must only be used as a property of a set. xbrlqe:InvalidObjectProperty

$A = set(a,b,c);
$B = set(c,d,e);
$A & $B

$A = set(a,b,c);
$B = set(c,d,e);
$A intersect $B

$A = set(a,b,c);
$B = set(c,d,e);
$A.intersect($B)

set(c)

When the intersect is used as a property of none a value of none will be returned.

$A = set(a,b,c);
none.intersect($A)

none

If the & or intersect operator is used with a value of none, the processor will return an error. xbrlqe:IncompatibleOperator

$A = set(a,b,c);
none & ($A)

9.1.1.3 difference(set to deduct) : set

The difference operation will produce the difference between two sets. The elements in the second set are deducted from the first set.

The difference of two sets can also be performed using the - operator.

A difference performed using the - operator must only be performed between two sets. A difference performed using the - operator between two incompatible collections results in a xbrlqe:IncompatibleOperator error.

A difference performed using the difference property will cast the list in the argument to a set. The difference property must only be used as a property of a set or a dictionary. xbrlqe:InvalidObjectProperty

$A = set(a,b,c);
$B = set(c,d,e);
$A - $B

$A = set(a,b,c);
$B = set(c,d,e);
$A.difference($B)

set(a,b)

When the difference is used as a property of none a value of none will be returned.

$A = set(a,b,c);
none.difference($A)

none

When the - operator is used to difference none with a set the value of the set will be returned. The - operator has special handling for missing data. If one side of the - operator is none then the value of the set is returned.

none - set(1,2,3)

9.1.1.4 symmetric-difference(sets to diff) : set

The symmetric-difference operation will produce a set of elements which are in either of the sets, but not in their intersection

The symmetric-difference of two sets can also be performed using the ^ operator.

A symmetric-difference performed using the ^ operator must only be performed between two sets. A symmetric-difference performed using the ^ operator between two incompatible collections results in a xbrlqe:IncompatibleOperator error.

$A = set(a,b,c);
$B = set(c,d,e);
$A ^ $B

$A = set(a,b,c);
$B = set(c,d,e);
$A.symmetric-difference($B)

set(a,b,d,e)

When symmetric-difference is used as a property of none a value of none will be returned.

$A = set(a,b,c);
none.symmetric-difference($A)

none

If the ^ operator is used with a value of none, the processor will return an error. xbrlqe:IncompatibleOperator

$A = set(a,b,c);
none ^ ($A)

9.1.1.5 is-subset(set to test) : boolean

The is-subset operation will produce a boolean result if one set is a subset of another set.

The is-subset of two sets can also be performed using the <= operator.

A is-subset performed using the <= operator must only be performed between two sets. A <= operator between two incompatible collections results in a xbrlqe:IncompatibleOperator error.

$A = set(a,b,c);
$B = set(a,b,c,d,e);
$A.is-subset($B)

$A = set(a,b,c);
$B = set(a,b,c,d,e);
$A <= ($B)

true

When is-subset is used as a property of none a value of none will be returned.

$A = set(a,b,c);
none.is-subset($A)

none

If the <= operator is used with a value of none, the processor will return none.

$A = set(a,b,c);
none <= ($A)

9.1.1.6 is-superset(set to test) : boolean

The is-superset operation will produce a boolean result if one set is a superset of another set.

The is-superset of two sets can also be performed using the >= operator.

A is-superset performed using the >= operator must only be performed between two sets. A >= operator between two incompatible collections results in a xbrlqe:IncompatibleOperator error.

$A = set(a,b,c);
$B = set(a,b,c,d,e);
$B.is-superset($A)

$A = set(a,b,c);
$B = set(a,b,c,d,e);
$B >= ($A)

true

When is-superset is used as a property of none a value of none will be returned.

$A = set(a,b,c);
none.is-superset($A)

none

If the >= operator is used with a value of none, the processor will return none.

$A = set(a,b,c);
none >= ($A)

9.1.2 Set Properties

Sets have the following properties:

• agg-to-dict (see Section 11.12.17)

• join (see Section 11.12.13)

• length (see Section 11.7.6)

• sort (see Section 11.12.14)

• to-list (see Section 11.12.16)

• to-dict (see Section 11.12.15)

• to-json (see Section 11.14.7)

9.2 Lists

A list is an ordered collection that can include duplicates. Lists can be sorted. A list is defined using the list() function.

9.2.1 List Properties

• agg-to-dict (see Section 11.12.17)

• index (see Section 11.13.1)

• join (see Section 11.12.13)

• length (see Section 11.7.6)

• sort (see Section 11.12.14)

• to-csv (see Section 11.13.2)

• to-dict (see Section 11.12.15)

• to-json (see Section 11.14.7)

• to-set (see Section 11.13.3)

9.2.2 Set and List Operators

9.3 Dictionaries

A dictionary is an unordered set of key: value pairs, with the requirement that the keys are unique (within one dictionary). A dictionary can be defined using the dict() function.

dict(list('AAxis','AMember'),list('BAxis','BMember'))

set(list('AAxis','AMember'), list('BAxis','BMember')).to-dict

A dictionary cannot be created using curly brackets.

9.3.1 Dictionary Specific Operations

9.3.1.1 Dictionary Index

To retrieve a value from a dictionary based on a key the square bracket operator [] is used.

$A = dict(list('AAxis','AMember'), list('BAxis','BMember'));
$A['AAxis']

'AMember'

The index operator of a dictionary will return none if the key value does not exist in the dictionary.

9.3.1.2 In Dictionary Operator

The in operator can be used to test if a key is in a dictionary. The equivalent property is has-key (see Section 11.14.2).

$A = dict(list('AAxis','AMember'), list('BAxis','BMember'));
$B = 'AAxis'
$B in $A

true

9.3.1.3 union(dictionary to union) : dictionary

The union operation will produce a union of two dictionaries, eliminating any duplicate keys contained in either dictionary. If the same key is added, then the add is not performed for the matching key.

The union of two dictionaries can also be performed using the + operator.

A union performed using the + operator must only be performed between two dictionaries. A union performed using the + operator between two incompatible collections results in a xbrlqe:IncompatibleOperator error.

$A = dict(list('AAxis','AMember'), list('BAxis','BMember'));
$B = dict(list('YAxis','YMember'));
$A + $B

$A = dict(list('AAxis','AMember'), list('BAxis','BMember'));
$B = dict(list('YAxis','YMember'));
$A.union($B)

dict(list('AAxis','AMember'), list('BAxis','BMember'), list('YAxis','YMember'));

When the union is used as a property of none a value of none will be returned.

none.union(dict(list('AAxis','AMember'), list('BAxis','BMember')))

When the + operator is used to union none with a dictionary the value of the dictionary will be returned. The + operator has special handling for missing data. If one side of the + operator is none then the value of the dictionary is returned.

none + dict(list('YAxis','YMember'))

dict(list('YAxis','YMember'))

9.3.1.4 difference(dictionary to deduct) : dictionary

The difference operation will produce the difference between two dictionaries. The dictionary pairs in the second dictionary are deducted from the first dictionary. The keys and values must match. If not then the left dictionary is maintained.

The difference of two dictionaries can also be performed using the - operator.

A difference performed using the - operator must only be performed between two dictionaries or a dictionary and list (see Section 9.3.1.5). A difference performed using the - operator between two incompatible collections results in a xbrlqe:IncompatibleOperator error.

The difference property must only be used as a property of a set or a dictionary. xbrlqe:InvalidObjectProperty

$A = dict(list('AAxis','AMember'), list('BAxis','BMember'));
$B = dict(list('BAxis','BMember'));
$A - $B

$A = dict(list('AAxis','AMember'), list('BAxis','BMember'));
$B = dict(list('BAxis','BMember'));
$A.difference($B)

dict(list('AAxis','AMember'))

When the difference is used as a property of none a value of none will be returned.

$A = dict(list('AAxis','AMember'), list('BAxis','BMember'));
none.difference($A)

none

When the - operator is used to difference none with a dictionary the value of the dictionary will be returned. The - operator has special handling for missing data. If one side of the - operator is none then the value of the dictionary is returned.

none - dict(list('AAxis','AMember'), list('BAxis','BMember'))

9.3.1.5 Subtract List of Keys Operator

The subtract dictionary key operator removes a key(s) and associated value from a dictionary. If keys don't match the left dictionary is maintained.

The removal or keys is performed using the - operator. On the left side is the dictionary to remove keys from and on the right side of the - operator is a list of keys to remove.

$A = dict(list('AAxis','AMember'), list('BAxis','BMember'));
$B = list('BAxis');
$A - $B

dict(list('AAxis','AMember'))

9.3.2 Dictionary Properties

• has-key (see Section 11.14.2)

• join (see Section 11.14.3)

• length (see Section 11.7.6)

• keys (see Section 11.14.4)

• values (see Section 11.14.5)

• to-json (see Section 11.14.7)

• to-spreadsheet (see Section 11.14.6)

10 Operators

Operators are used between variables, collections and values.

10.1 Comparison Operators

Comparison operators compare two values for equivalence. The result is a boolean value.

10.2 Boolean Operators

10.3 Numeric Operators

A division by zero will produce a divide by zero error xbrlqe:DivByZero.

10.4 Using Comparison Operators with Time

10.5 Order of Operator Precedence